The goal of this work is to identify mouse receptors that are ideally
specific, or at least enriched in mMO germ cells
In the analysis below, we:
Read in the
mMO151_MFTAv2.2_mouse_final.RDS and either germ = Yes or germ = No
Attempted to run DGE analysis (fit_models()) on a downsampled atlas +
entire mMO cds However, this led to memory issues. Moving forward I ran
DGE analysis using just the mMO subset. The DGE results were filtered to
only considered genes that are present in the mouse receoptors list.
Finally, results were filtered by lFC (estimate) and q_value (corrected
p-value) and plotted for visualization.
We find there are two datasets in this cds, Invitro and MFTAv2.2. They are well mixed asside from immune / endothelial cell calls.
To identify germ enriched genes. We first define a variable “germ”
and assign cells either Yes or No Depending on whether their
assigned_cell_type / partition belongs to PGC/germ.
Now we run DGE analysis. Attempts to include a down-sampled portion of
the MFTAv2.2 and run DGE analysis on a combined object continually hit
memory limit issues. However, because the datasets are well mixed, and
becasue the InVitro subset itself has thousands of cells it is
reasonable to run DGE analysis on just the InVitro subset.
After DGE analysis, we filter by lfc >3 and q_value < 0.05.
Next, we subset the results to only consider genes identified as mouse
receptors. Finally, we plot the results.
We can see that Tex101 is a good candidate for a receptor/surface marker
that is predominantly expressed in the PGC/Germ compartment. Other
candidates have similar enrichment/specificity, but much lower relative
expression.